This could possibly be an example of reverse causality, as we cannot rule out that the suPAR level was already increased before the purchase of antidepressants. For men, the effect was greater when the purchase was temporally closer to the suPAR measurement. For women, the effect was independent of purchase timing. There was, however, a trend towards a decreased strength of the association when the time between suPAR measurement and the purchase of the antidepressant medicine increased. The timing categories were based on clinical guidelines for treatment in the acute and maintenance phases and are in accordance with prior studies. Our findings are in line with studies of depression and CRP that found an association between CRP levels and prior depression. Two of these studies, however, only detected the association in men and not in women. In this study, we found an effect for both men and women. It is well known that there is an increased prevalence of depression among patients with diseases such as diabetes, myocardial infarction and cancer. Thus, to examine whether the relationship between suPAR levels and the subsequent use of antidepressants was caused by another underlying disease affecting both suPAR level and depression, we adjusted the analyses for prescription medications other than antidepressants. The purchase of medication, other than antidepressants, was not associated with an increased risk of antidepressant usage and did not affect the relationship between suPAR levels and the incident use of antidepressants. This does not suggests that treatment with antidepressant medicine could have been initiated in connection with an underlying other disease. This study is for the outcome based on Paclitaxel register data using the purchase of antidepressant medicine as a proxy for depression. The same ATC code has previously been used in other population-based studies of depression. Whether our use of register data on antidepressant use captures the donor with a depression is subject to discussion. However in Denmark can such medication only be prescribed by physicians and regulation and surveillance of the use of medication is quite strict. The Table 2. sensitivity of antidepressant use as a proxy for major depression has been evaluated to be up to 50% while, more importantly, specificity seems to be more than 90%. Thus, it seems reasonable to use the purchase of antidepressant medicine as a proxy for depression, acknowledging that milder forms of depression may not all be recorded. The sensitivity of 50% indicates that many people with depression are not prescribed medication and there can be a long delay between the onset of depression and initiation of the antidepressant use. Further, antidepressants are used for a number of indications other than major depression. Thus, we cannot completely rule out that a proportion of participants who started using antidepressants after the suPAR measurement were already depressed or had been depressed before blood collection without receiving an antidepressant at that time. Also did we in this analysis exclude all donors who had purchased antidepressants prior to the suPAR measurement. However older donors could potentially have been treated with antidepressants before the register started in 1995. A limitation of our study is the lack of information on relevant potential confounders such as smoking, alcohol consumption, BMI and physical activity. Smoking may lead to increased suPAR levels and is common among depressed patients. Obesity has been suggested to be a potential link between depression and elevated CRP ; however, no effect of obesity on depression was found in the 12-year US national health survey.